On Wed, Feb 10, 2010 at 3:29 AM, Bryce Nesbitt <bryce2(at)obviously(dot)com> wrote:
> Or, if you want to actually read that query plan, try:
> http://explain.depesz.com/s/qYq
Much better, though I prefer a text attachment... anyhow, I think the
root of the problem may be that both of the subquery scans under the
append node are seeing hundreds of times more rows than they're
expecting, which is causing the planner to choose nested loops higher
up that it otherwise might have preferred to implement in some other
way. I'm not quite sure why, though.
...Robert